The past several years have witnessed a notable uptick in workplace artificial intelligence related legislation and agency enforcement attention, specifically focused on the infusion of AI or so-called automated decision-making tools.
Colorado’s new Artificial Intelligence Act, for example, designates employment as a “high-risk” sector of AI applications and has heightened concerns of lawmakers and corporate executives. Lawsuits, such as Mobley v. Workday and Moffatt v. Air Canada, underscore the concerns of employment candidate screening, recruitment and conversational AI. Most recently, the US Equal Employment Opportunity Commission issued a Determination finding cause to believe the employer violated the Older Workers Benefit Act by using AI in a reduction in force that adversely impacted older workers. A complaint in the Southern District of New York against IBM and its spinoff technology company, Kyndryl, promptly followed.
Perhaps not surprisingly, over the past few years, the State of New York (“NYS”), following the lead of New York City, has introduced several bills that would regulate the use of AI infused decision-making tools. One such bill, called New York Workforce Stabilization Act (“NYWFSA”) was introduced in May 2024 by Senators Michelle Hinchey and Kristen Gonzalez. They will likely re-introduce the NYWFSA during the upcoming January 2025 legislative session intending to “stabilize” New York’s labor market at a time when the deployment of AI may fundamentally alter the New York industrial landscape.
At around the same time as the New York legislature was considering the NYWFSA, McKinsey & Company published a comprehensive review titled, “Generative AI and the Future of New York” articulating challenges for New York employers and employees resulting from the deployment of Generative AI. The report estimates that within roughly five years, there will be as many as 1.1 million “occupational shifts” in the New York region, and one-third or approximately 380,000 of these shifts will be the result of Gen AI. Bottom line is that AI affects a “far broader range of occupations than existing automation technologies do.” Thus, while AI will undoubtedly benefit work efficiencies over a great swath of jobs, job transitions and displacement are likely to occur.
In its current form, the NYWFSA appears be New York’s attempt to ameliorate the impending impact of such AI disruptions, by gathering information about how New York employers with one hundred or more state-based employees will use AI, and its impact on New York workers. It will thus require covered employers to undertake AI assessments prior to deployment and provide the result of such assessments to the NYS Department of Labor. NYWFSA also sets forth surcharges on employers displacing employees with AI and using the proceeds to furnish worker retraining, among other things.
As noted above, NYWFSA would not be the first workplace AI legislation in New York. New York City’s Local Law 144 (“Local Law 144”) effectively commenced New York’s regulation of AI in businesses in 2023. As we have previously written about Local Law 144, from its passage, implementation, and FAQs, to last year’s update, Local Law 144 regulates employers and employment agencies that use automated employment decision tools. The law provides certain applicants and employees a private right of action to compel enforcement of the law’s provisions as well as significant monetary penalties for noncompliance.
I. Impact Assessments
Under the NYWFSA, covered employers would be required to provide the NYS Department of Labor with a comprehensive impact assessment of the AI before deployment and undertake impact assessments every two years thereafter.
- The impact assessment would include:
- information on the objectives intended for deploying the AI tool or system;
- a summary of the underlying core algorithms;
- training data utilized for the algorithmic model;
- personal or sensitive data was used to train the model; and,
- the protocols or mechanisms for data storage.
II. Surcharges
NYWFSA establishes a surcharge on certain corporations that use AI, conduct data mining, or have more than fifteen employees displaced by AI at a rate of 2% of the corporation’s business income base. The bill delineates two categories of surcharges. One, the bill imposes a “worker displacement surcharge” on “corporations that terminate the employment or substantially reduce the hours of [15] or more employees due to any system or process that uses algorithms, computational models, AI techniques, robotic hardware, or a combination thereof to automate, support, or replace human labor.” Additionally, the bill imposes a “data mining surcharge” for using “a process involving pattern-based queries, searches, and other analyses of one or more electronic databases.” Surcharge revenues will go to the Department of Labor’s worker retraining programs, workforce development programs, and the unemployment insurance fund.
Agency Enforcement
If the New York legislature passes NYWFSA, the Commissioner and the New York Department of Labor act as NYWFSA’s primary collectors of impact assessments and surcharges. Companies must pay worker displacement surcharges to the Commissioner annually for deposit into banks designated by the NY Comptroller. In contrast, administration and collection of data mining surcharges will be like the collection of taxes. The bill, as of this draft, does not provide punitive provisions for violators and it is undetermined which agency will enforce the bill’s requirements. It appears, however, from recent enforcement actions from various state and federal government agencies that remedial activities will increase as the deployment of workplace AI continues apace.
Implications for Employers
Given the legislative and regulatory intentions of NYS and an increasing number of other states and local jurisdictions, as well as federal and state government regulatory agencies, an employer’s consideration of deploying AI to increase workplace efficiencies should be accompanied by comprehensive review and analysis of its contemplated training data for the algorithmic model (input) and the impact of the output.
For starters, employers should review their current use of AI processes or tools to determine whether the technologies meet current business needs, are job related and consistent with business necessity, and compliant with applicable laws and regulations. Employers considering AI should undertake an assessment of the areas in which the company may most benefit from workplace AI, e.g., candidate sourcing; resume screening; hiring and onboarding; candidate testing; determining employee career paths; succession planning, etc. This would require consulting with AI vendors to ensure comprehension of the system’s considerations of bias, reviewing contracts with AI vendors to assess risks and liabilities, and preparing communication to inform job applicants and employees about AI’s involvement in decision-making processes.
NYWFSA has not provided parameters for acceptable AI usages and unacceptable AI usages. The New York Department of Labor, based on proposed legislation governing AI usage in New York agencies may analyze for accuracy, fairness, bias and discrimination, cybersecurity vulnerabilities, privacy risks, and foreseeable misuse.
In the interim, as we await further guidance, employers are encouraged to make sure that their uses of workplace AI tools are consistent with general principles to minimize the risks of AI implementation while also maximizing the benefits of its use. Employers should ensure that data collection is handled and secured responsibly, that AI systems are designed and deployed to safeguard workers, and that AI use respects workers’ rights. Employers are advised to ensure that workers are informed and properly trained in AI use, and that such uses of AI are meant to enhance job quality for workers. Creating and establishing clear and communicated procedures on workplace AI use and oversight is pivotal to ensure that this new feature in the workplace is utilized in the best and most productive way.
And the Beat Goes On….
Legislative trends in NYS prove AI acceleration will grow in conjunction with AI regulation. At the start of the year, Governor Kathy Hochul launched the Empire AI Consortium, an artificial intelligence computing center facilitating statewide innovation, research, and development of AI technologies. Governor Hochul noted, perhaps prophetically, that whoever dominates this next era of AI will dominate history and the future. It is no coincidence that New York senators introduced NYWFSA to pioneer New York’s first AI workforce regulation in the same year. States like New York want to be at the forefront of this next era of AI while protecting and preparing the labor market. The trend does not stop with state actions. In the U.S. Congress, a bipartisan pair of senators introduced the Small Business Artificial Intelligence Training and Toolkit Act to accelerate small business use of artificial intelligence. Previously, President Biden released the Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, as discussed here. In the same regulatory vein, Majority Leader Chuck Schumer (NY-D) continues to host AI Insight Forums, involving AI experts, civil rights advocates, and industry members to inform Congress’ work on regulating AI. Most recently, the Bipartisan Senate AI Working Group released A Roadmap for Artificial Intelligence Policy in The United States Senate. The push for AI regulation will only increase as we continue to use AI technologies.
Epstein Becker Green’s artificial intelligence and cybersecurity and data privacy and management attorneys can assist in providing guidance with respect to any aspect of the infusion of AI in the workplace and/or consumer facing utilization. To the extent that you have any questions on this post or otherwise would like further information, any one of the authors would be happy to assist.
Janae Barrett (Law Clerk - Admission Pending) and América Garza (Law Clerk – Admission Pending) in the firm’s New York office, contributed to the preparation of this article.
